README.md

magrittr - Ceci n'est pas un pipe.

Introduction

This is an experimental redesign of magrittr built on the principle of functional sequences.

For most practical purposes, it works as usual:

value %>%
  fun1(arg1) %>%
  fun2(arg2, .) %>%
  fun3

The tee operator is there:

value %T>%
  plot %>%          # usually doesn't return anything
  some_function     # receives value

And the compound assignment operator:

iris$Sepal.Length %<>% multiply_by(2)

The main difference is that the entire right-hand side is "compiled" into a functional sequence (a function which applies the individual right-hand sides sequentially). When the left-hand side is a value, the result of applying this function is returned. One can also use the magrittr placeholder, the . as left-hand side, in which case functional sequences itself is returned:

fun <- . %>% cos %>% sin %>% sum(na.rm = TRUE)

The function list can be extracted as

functions(fun)

For easy inspection these have their own print method:

> fun
Functional sequence with the following components:

 1. cos(.)
 2. sin(.)
 3. sum(., na.rm = TRUE)

Use 'functions' to extract the individual functions.

The one-time "compilation" of the functional sequence is also useful in repetitive situations:

loong_vector %>% lapply(. %>% a %>% b %>% c %>% d)

In this case the expressions a--d only has to be determined once.

Relevant other differences

• There is currently no lambda function in this version.
• magrittr does not try to maintain a call as previously.